THE LAST PLANNER SYSTEM OF PRODUCTION CONTROL by
HERMAN GLENN BALLARD
A thesis submitted to the Faculty of Engineering of The University of Birmingham
for the degree of DOCTOR OF PHILOSOPHY
School of Civil Engineering Faculty of Engineering
ACKNOWLEDGEMENTS
Many individuals and companies contributed to this research. To name a few:
q David Seymour, my thesis advisor
q Anne Seymour, for giving me a home in Birmingham
q David Hoare and Peter Deasley, thesis examiners
q Greg Howell, business partner and co-thinker
q Lauri Koskela, for his example and inspiration
q Todd Zabelle and the Pacific Contracting team for their willingness to try new ideas
q Leo Linbeck III, Ed Beck and Kathy Jones of Linbeck Construction for sharing opportunity and data (3 of the 5 cases were Linbeck projects)
q Norm Barnes and the Barnes Construction team for access to projects
q Iris Tommelein, external thesis advisor and colleague at UC Berkeley
q Jeanne Ballard, my wife, for putting up with me, especially my absences from home
ABSTRACT
Project controls have traditionally been focused on after-the-fact detection of variances. This thesis proposes a control system, the Last Planner system, that causes the
realization of plans, and thus supplements project management's concern for management of contracts with the management of production.
The Last Planner system has previously been successively applied by firms with direct responsibility for production management; e.g., speciality contractors. This thesis extends system application to those coordinating specialists, both in design and construction, through a series of case studies, one of which also explores the limits on unilateral implementation by specialists.
In addition to the extended application, two questions drive this research. The first question is 1) What can be done by way of tools provided and improved implementation of the Last Planner system of production control to increase plan reliability above the 70% PPC level? Previous research revealed substantial improvement in productivity for those who improved plan reliability to the 70% level, consequently there is reason to hope for further improvement, possibly in all performance dimensions, especially with application across an entire project rather than limited to individual speciality firms. That question is explored in three case studies, the last of which achieves the 90% target.
The second research question is 2) How/Can Last Planner be successfully applied to increase plan reliability during design processes1? That question is explored in an extensive case study, which significantly contributes to understanding the design process from the perspective of active control, but unfortunately does not fully answer the question, primarily because the project was aborted prior to start of construction. However, it is argued that the Last Planner system is especially appropriate for design production control because of the value-generating nature of design, which renders ineffective traditional techniques such as detailed front end planning and control through after-the-fact detection of variances.
Issues for future research are proposed, including root cause analysis of plan failures and quantification of the benefits of increased plan reliability for both design and construction processes.
TABLE OF CONTENTS
Page Title Page Acknowledgements Abstract Table of Contents List of Figures List of Tables 1.0 Introduction 1-1 1.1 Conceptual Framework 1-1 1.2 Assumptions 1-3 1.3 Contribution to Knowledge 1-51.4 The Author's Role in the Research 1-5
1.5 Structure of the Dissertation 1-8
2.0 Critique of Production Control 2-1
2.1 What is Production Control? 2-1
2.1.1 The Meaning of “Production” 2-1
2.1.2 The Meaning of “Control” 2-3
2.2 Project Management 2-5
2.2.1 The Project Management Body of Knowledge 2-5 2.2.2 Critique of the Traditional Project Control Model 2-7 2.3 Previous Application of Production Control Concepts to the AEC Industry 2-12
2.3.1 Melles and Wamelink 2-12
2.3.2 Koskela 2-14
2.4 Criteria for a Production Control System 2-15
3.0 Description and History of the Last Planner System of Production Control 3-1
3.1 Hierarchical Structure 3-1
3.2 Should-Can-Will-Did 3-1
3.3 Production Unit Control 3-3
3.4 Work Flow Control 3-5
3.4.1 Constraints Analysis 3-9
3.4.2 Pulling 3-11
3.4.3 Matching Load and Capacity 3-14
3.7 Evaluation of Last Planner against Criteria for Production Control Systems 3-24 3.8 Research Questions: 1) What can be done by way of tools provided and 3-25
improved implementation of the Last Planner system of production
control to increase plan reliability as measured by Percent Plan Complete? 2) How/Can Last Planner be successfully applied to increase plan reliability during design processes?
4.0 Research Methodology 4-1
4.1 Introduction 4-1
4.1.1 Engineering Management as a Field of Study 4-1 4.1.2 Competing Engineering Management Paradigms 4-4
4.2 Research Design 4-6
4.2.1 Research Question 4-6
4.2.2 Research Strategies 4-8
4.3 Research Methods 4-9
4.3.1 Data Collection 4-9
4.3.2 Data Analysis and Evaluation 4-10
4.3.3 Case Studies 4-12
5.0 Case One: CCSR Project 5-1
5.1 Project Description and Last Planner Implementation 5-1
5.2 PPC and Reasons 5-2
5.3 Observations 5-10
5.4 Learnings 5-11
6.0 Case Two: Next Stage 6-1
6.1 Project Description and Last Planner Implementation 6-1
6.2 Data 6-2
6.2.1 PPC and Reasons 6-2
6.2.2 Observations 6-5
6.2.3 Feedback from participants 6-5
6.3 The Nature of the Design Process and Implications for Process Control 6-8
6.4 Evaluation of Last Planner Implementation 6-10
6.5 Learnings 6-12
7.0 Case Three: Pacific Contracting 7-1
7.1 Project Description and Last Planner Implementation 7-1
7.2 PPC and Reasons 7-2
7.3 Observations 7-7
7.4 Learnings 7-8
8.0 Case Four: Old Chemistry Building Renovation 8-1
8.1 Project Description and Last Planner Implementation 8-1
8.2 PPC and Reasons 8-2
8.3 Observations 8-4
9.0 Case Five: Zeneca 9-1 9.1 Project Description and Last Planner Implementation 9-1
9.2 PPC and Reasons 9-1
9.3 Constraint Analysis and Make Ready 9-3
9.4 Observations 9-6
9.5 Learnings 9-6
10.0 Conclusions 10-1
10.1 Summary of Case Study Results 10-1
10.2 Research Question: What can be done by way of tools provided and 10-2 improved implementation of the Last Planner system of production
control to increase plan reliability as measured by Percent Plan Complete?
10.3 Research Question: How/Can Last Planner be successfully applied to 10-3 increase plan reliability during design processes?
10.4 Directions for Future Research 10-4
10.5 Conclusions 10-9
Glossary of Terms G-1
List of References R-1
Bibliography Biblio-1
Appendix A: Next Stage Kickoff Meeting A-1
Appendix B: Next StageTeleconferences B-1
Appendix C: Next Stage Action Items Log C-1
Appendix D: Next Stage Issues Log D-1
LIST OF FIGURES
Page 3.1 The formation of assignments in the Last Planner System 3-2
3.2 Lookahead Process 3-6
3.3 Make Ready by Screening and Pulling 3-11
3.4 A Traditional (Push) Planning System 3-13
3.5 Last Planner-A Pull System 3-14
3.6 The Last Planner System 3-16
3.7 PPC (Nokia Project) 3-23
3.8 Participant Survey (Nokia Project) 3-23
5.1 CCSR-Weekly PPC 5-5
5.2 CCSR-Reasons for Noncompletion 5-6
5.3 CCSR-PPC without rain 5-6
6.1 Next Stage-PPC 6-4
7.1 Pacific Contracting-PPC 7-2
7.2 Pacific Contracting-Reasons 7-7
8.1 Old Chemistry Building-PPC 8-3
8.2 Old Chemistry Building-Reasons for Noncompletions 8-4
9.1 Zeneca-PPC 9-2
9.2 Zeneca-Reasons 9-2
10.1 Activity Definition Model 10-4
10.2 Reasons Analysis Hierarchy-Directives 10-6
10.3 Reasons Analysis Hierarchy-Prerequisites 10-7
10.4 Reasons Analysis Hierarchy-Resources 10-8
LIST OF TABLES
Page
1.1 Conversion/Flow/Value 1-1
3.1 Functions of the Lookahead Process 3-7
3.2 Construction Lookahead Schedule 3-9
3.3 Engineering Lookahead Schedule 3-10
3.4 Constraints Analysis 3-12
5.1 CCSR-Weekly Planning Cycle 5-3
5.2 CCSR-Constraints Analysis Form 5-4
5.3 CCSR-PPC and Reasons Data 5-5
5.4 CCSR-Reasons for Noncompletion (detailed and categorized) 5-7
6.1 Next Stage-Reasons for Noncompletion 6-2
6.2 Next Stage-PPC Data 6-3
6.3 Next Stage-Reasons 6-5
7.1 Pacific Contracting-PPC Data and Reasons 7-3
8.1 Old Chemistry Building-PPC Data 8-3
CHAPTER ONE: INTRODUCTION
1.0 Conceptual Framework
Production processes can be conceived in at least three different ways: 1) as a process of converting inputs to outputs, 2) as a flow of materials and information through time and space, and 3) as a process for generating value for customers. All three conceptions are appropriate and necessary. However, the conversion model has been dominant in the AEC (architectural/engineering/construction) industry until very recently (Koskela and Huovila, 1997).
Table 1.1
Conversion View Flow View Value Generation
Nature of Construction a series of activities which convert inputs to outputs.
the flows of information & resources, which release work: composed
of conversion, inspection, moving and
waiting. a value creating process which defines and meets customer requirements. Main Principles Hierarchical decomposition of activities; control and
optimization by activity. Decomposition at joints. Elimination of waste (unnecessary activities), time reduction. Elimination of value loss - the gap between achieved and possible value.
Methods & Practices
Work breakdown structure, critical path
method. Planning concerned with timing start and responsibility for activities through contracting or
assigning.
Team approach, rapid reduction of uncertainty, shielding, balancing, decoupling.
Planning concerned with timing, quality and
release of work.
Development and testing of ends against means to determine requirements. Planning concerned with work structure, process and participation. Practical Contribution Taking care to do necessary things.
Taking care that the unnecessary
is done as little as possible.
Taking care that customer requirements are
met in the best possible manner.
.
Conversion/Flow/Value2
The design and construction of AEC facilities (buildings, process plants, airport terminals, highways, etc.) poses difficult management problems to which the models and techniques based on the conversion view have proven inadequate. Tradeoffs between competing design criteria must be made throughout the design process, often with incomplete information and under intense budget and schedule pressure. Increasingly, projects are subject to uncertainty because of the pace of technological change and the rapid shifting of market opportunities and competitor actions.
Production management concepts and techniques based on the conversion model have not proven capable of solving these difficult problems. The heart of the conversion model is the assumption that the work to be done can be divided into parts and managed as if those parts were independent one from another. Management techniques such as work breakdown structures and earned value analysis belong to this conversion model. Work breakdown structures are driven by scoping and budget concerns and have the objectives of insuring that all the work scope is included in one of the parts, insuring that no work scopes overlap, and allocating costs to each part such that the rollup yields the total for the project. This division into parts is necessary in order to allocate responsibility to internal or external work centers, which can subsequently be controlled against scope, budget, and schedule commitments.
This is fundamentally a contracting mentality, which facilitates the management of contracts rather than the management of production or work flow. Production management is the ‘local’ responsibility of those to whom the various parts are assigned or contracted. If everyone meets their contractual obligations, the project performs successfully. Unfortunately, this approach is the opposite of robust. When something
If a management philosophy and tools are needed that fully integrate the conversion, flow, and value models, we might consider the product development processes employed by firms designing and manufacturing consumer products (automobiles, printers, toasters, etc.). Such processes have developed potentially useful concepts especially in the area of value; identification of customer needs and translation into engineering specifications (Ulrich and Eppinger 1993). Product development processes also are struggling with other issues relevant to the design of AEC facilities, including design decomposition, organizational means for integration, etc. (Hayes, et al, 1988; Eppinger, et al, 1990; Gebala and Eppinger, 1991).
As a contribution to the integration of all three models, this thesis applies the flow model to managing the design and construction of AEC facilities. Conceptualizing the design and construction process as a flow of information and materials lends itself to reducing waste by minimizing time information or materials spend waiting to be used, time spent inspecting information or materials for conformance to requirements, time spent reworking information or materials to achieve conformance, and time spent moving information or materials from one specialist to the next. Further, conceptualizing the design and construction process as a flow of information and materials allows coordination of interdependent flows and the integration of design with supply and site construction.
1.2 Assumptions
q Current construction industry production management thinking and practice is dominated by the conversion model, consequently value generation and flow management concepts and techniques are underdeveloped.
q To be consistent with all three models, conversion, flow, and value, production management should be conceived as having the purpose of creating customer value while minimizing waste in time and cost. “Customer value” is understood to include not only the fitness for use of the facility considered with regard to functionality, but also with regard to all other criteria to which the customer attaches value, e.g., project delivery within a time and for a cost that meets the customer’s market and financial needs.
q "Production" is understood to include both designing and making. The historical development of production theory in manufacturing has erroneously suggested that production is entirely concerned with 'making'.3
q Production management is conceived to consist of criteria determination and work structuring in the ‘planning’ phase, and to consist of work flow control and
production unit control in the ‘execution’ or ‘control’ phase.
This thesis treats only control functions, not planning functions. It does not treat the very first and fundamental production management activity; i.e., the determination of
customer needs and their translation into design criteria. Criteria determination belongs to the value generation view. This thesis treats only the flow view. Similarly, work structuring activities such as identification, sequencing, and scheduling tasks are also not
3
There may be differences between the U.S. and U.K. in the use of these terms. Hence the effort to be precise. For the most part, the theory of producing artifacts has emerged
treated here. The scope of this thesis is the control functions of production unit control and work flow control.
1.3 Contribution to Knowledge
This dissertation proposes to make the following contributions to knowledge:
q Adapted from manufacturing4, a system for production control, the Last Planner system, is presented that exemplifies the concept of control as causing events to conform to plan, as distinct from the traditional conception of project control in terms of after-the-fact variance detection.
q Appropriate application of the production control system is shown to improve work flow reliability, which promises substantial benefits in project cost and duration reduction.
q Improvements to the Last Planner system of production control are developed and tested in a series of case studies, resulting in new concepts and techniques.
Project controls in the AEC industry have focused on detecting variances from project objectives for cost and schedule, and have not directly dealt with the management of production. The Last Planner system of production control has proven an effective tool for improving the productivity of the production units that implement its procedures and techniques (Ballard and Howell, 1997). This dissertation shifts the focus from the productivity of the immediate production unit to the reliability of work flow between production units, and also extends application of the system to design.
1.4 The Author's Role in the Research
4
The Last Planner system has been in development by the author since 1992. Several papers have previously been published by this author on the subject, the first of them in 1993 (Ballard, 1993) at the founding conference of the International Group for Lean Construction. Last Planner research began with a focus on improving the quality of assignments in weekly work plans (Koch Refinery Mid-Plants Project, 1993-45), added a lookahead process to shape and control work flow (PARC, 19956; DMOS-6, 19967), and eventually was extended from construction to design (Nokia8 and Hewlett-Packard9, 1996). During that development, the objective shifted from improving productivity to improving the reliability of work flow. This resulted from a change in conceptual framework. The initial framework came from the quality management and productivity improvement initiatives that dominated construction industry performance improvement efforts in the 1980s. The shift to work flow reliability reflected the author's increasing awareness of the revolution in manufacturing inspired by the Toyota Production System and eventually labeled "lean production", and also contact with the thinking of Lauri Koskela regarding production theory and its application to the construction industry.
A key metric of the Last Planner system is the percentage of assignments completed (PPC), which is clearly a defect rate and a product of the quality management mentality. Given the objective of improving productivity, measurements were made of the relationship between the defect rate of a crew, its PPC, and the productivity of that crew. Not surprisingly, such measurements revealed a positive correlation10. However, the
5
Ballard and Howell, 1997
6
Ballard, Howell, and Casten (1996)
7
Ballard and Howell, 1997
8
Koskela, Ballard, and Tanhuanpaa (1997)
activity focus characteristic of the productivity improvement 'mind' concealed the importance of that crew's PPC for the productivity of the crews that followed it and built upon its work product. Even the introduction of a lookahead process was motivated initially by the observation that simply shielding a crew from poor assignments was insufficient to optimize crew productivity. To do so required matching load and capacity, both of which required managing load or work flow. The more powerful and fundamental opportunity to coordinate action among multiple crews was hidden by the dominance of what Koskela has called the "conversion model" and its exclusive focus on the activity as the unit of control rather than work flow.
Prior to the founding of the Lean Construction Institute (LCI) in August of 199711, the Last Planner system had evolved to roughly its current form, with a clear conceptual basis in production theory a la Koskela and an explicit and self-conscious objective of managing work flow. What remained to be done was to learn how to improve work flow reliability above the 35%-65% range commonly discovered up to that time. One purpose of this dissertation is to describe what was done to improve work flow reliability, measured by PPC, and the results achieved. That improving work flow reliability is beneficial hardly requires argument. However, identifying and quantifying the specific benefits will be a matter for future research. The second purpose of this research is to explore applicability of the Last Planner system to design.
11
The Lean Construction Institute was founded in August of 1997 as a partnership between Gregory A. Howell and Glenn Ballard, dedicated to research, training and consulting in construction industry production management. Subsequently, Iris Tommelein and Todd Zabelle have become partners in the enterprise, along with Mark Reynolds, Managing Director of Lean Construction International, based in London. All the case studies reported in this thesis were undertaken as research projects for LCI, of which this author is Research Director. All case studies were carried out under the
1.4 Structure of the Dissertation
Traditional project control theory and practice is described and critiqued in Chapter Two. The Last Planner System of Production Control is presented in Chapter Three as satisfying the requirements revealed by the critique. Chapter Four describes the research methodology used in the dissertation and is followed by Chapters 5, 6, 7, 8, and 9, each devoted to a case study. Conclusions from the case studies are reported in Chapter 10, followed by a glossary of terms, a list of references, a bibliography, and an appendix consisting of documents from the design case, Next Stage.
CHAPTER TWO: CRITIQUE OF PRODUCTION
CONTROL
2.1 What is Production Control?
The purpose of this chapter is to provide a critique of production control theory and practice. But first it is necessary to clarify what is meant by “production control”.
2.1.1 The Meaning of “Production”
Production has been an explicit topic of study primarily in industrial engineering, which has dealt almost entirely with one type of production; namely, manufacturing (in the sense of 'making'), with only occasional forays into construction, plant maintenance, building maintenance, agriculture, forestry, mining, fishing, etc. Design and engineering have infrequently been conceived as production processes; the focus almost entirely being placed on making things rather than designing them.
Although the meaning of the term at its most universal is synonymous with “making”, “manufacturing” is most commonly12 used to denote the making of many copies from a single design, and consequently is primarily focused on products for a mass market, most of those products being moveable from the place manufactured to the place of use. There are exceptions to the products being moveable, although still copies from a single design; e.g., ships and airplanes. Within the world of construction, manufacturing in this sense is approached mostly closely by 'manufactured housing'.
12
Exceptions occur with thinkers and writings regarding product development, which by its nature must integrate designing and making, at least in the sense of making prototypes.
Various types of making have been proposed, among them ‘assembly’, the joining of parts into a whole, as distinct from ‘fabricating’, the shaping of materials. For example, construction is often categorized as a type of ‘fixed position manufacturing’ (Schmenner, 1993), along with shipbuilding and airplane assembly. In all these instances of assembly, the assembled product eventually becomes too large to be moved through assembly stations, so the stations (work crews) must be moved through them, adding additional components and subassemblies until the artifact (building, bridge, tunnel, plant, house, highway, etc) is completed.
Many publications exist on the topic of production management in manufacturing, the larger part of which adopt the perspective of the industrial or production engineer (Bertrand et al, 1990; Hopp and Spearman, 1996; Murrill, 1991; Vollman et al, 1992). A subset of this category concern themselves with the psychological/sociological aspects of manufacturing management (Scherer, 1998). The development of alternatives to mass production over the last 40 years has been revolutionary. Early and influential production management theorists include Jack Burbidge (1983; 1988) and W. Edwards Deming (1986), to mention but a few from the West. Taiichi Ohno (1988) and Shigeo Shingo (1988) were the primary architects of the Toyota Production System, the archetype for lean production, so named in part to counterpose it to "mass" production. Burbidge's groundbreaking thought began to emerge in the 1960s. Deming was instrumental in the implementation of quality management and statistical quality control concepts and techniques in Japan after the 2nd World War. The work of Ohno and Shingo was concentrated in the period of the late 50's into the 70's. The Machine That Changed the World (Womack et al., 1990) reported the findings of an international study of the
which presented the principles and basic concepts behind the new forms of manufacturing and proposed to extend them to the entire enterprise. Womack and Jones have popularized and made more easily accessible the concepts and techniques of lean production.
Defining production as the designing and making of artifacts allows us to understand how construction is a type of production and also that design is an essential component in production generally and in construction specifically. Lauri Koskela (Koskela 1992, 1999; Koskela and Huovila 1997; Koskela et al. 1996, 1997) is the foremost production theorist in construction. His study of the applicability of newly emergent manufacturing concepts and techniques to the construction industry has driven him back to the development of a theory of production as such (Koskela, 1999).
2.1.2 THE MEANING OF “CONTROL”
The term “control” has a wide range of meanings. According to the Concise Oxford Dictionary, its meanings include to dominate, command; to check, verify; to regulate. It has long been associated with accounting. The Old French contreroller: to keep a roll of accounts.
Accounting is the essence of project control theory, more fully described in section 2.2.2 below (Diekmann and Thrush, 1986; Project Management Body of Knowledge (PMBOK), 1996; Riggs, 1986). The essential activity is monitoring actual costs or schedule performance against target in order to identify negative variances. Corrective action is obviously necessary in order to correct such negative variances, but the literature hardly addresses corrective action.
Industrial process control introduces feedback and feedforward mechanisms for regulating a process (Murrill, 1993). Feedback is initiated by a comparison of actual with target outputs. Feedforward is initiated by a comparison of actual with target inputs.
The artificial intelligence community contributes the blackboard system of control, in which coordination of a number of interdependent specialists is managed by rules for taking turns 'writing on a blackboard'; i.e., for contributing to their collaborative work (Hayes-Roth, 1985). AI adherents have been in the forefront of empirical study of design, and despite their technological orientation, have found social and organizational issues to be of great importance. Finger et al (1995) conclude: “The social process plays a major role in the articulation and realization of the product design, particularly in large projects.” (p.89). Bucciarelli (1984) reports that designers spend 85-90% of their time talking, writing, negotiating, meeting, searching, etc. as opposed to drawing and calculating.
Production control theorists working in manufacturing distinguish two primary ways of regulating work flow in manufacturing systems: push and pull. Push systems release materials or information into a system based on preassigned due dates (from a master production schedule, for example) for the products of which they are parts. Pull systems release materials or information into a system based on the state of the system (the amount of work in process, the quality of available assignments, etc) in addition to due dates (Hopp and Spearman, 1996). In factory systems, pull may be derivative ultimately from customer orders. In construction, pull is ultimately derivative from target completion dates, but specifically applies to the internal customer of each process. Applicability of these concepts to production control has been explored by this author
Some theorists (Kelly, 1994) propose that complex, dynamic systems are regulated not by anything resembling a central mind, but through the independent action of distributed decision makers. The following excerpt from Eric Scherer’s introduction to Shop Floor Control-A Systems Perspective indicates the emergence of a new conceptual framework,
“To master the challenges of the future, there must be a change in our thinking paradigm. Manufacturing is not deterministic! …the problem of systems design for shop floor control is no longer the problem of ‘optimization’. The reductionistic paradigm … needs to be replaced by a holistic paradigm of agile activity, dynamic behavior, and evolutionary development.”
2.2 Project Management
2.2.1 THE PROJECT MANAGEMENT BODY OF KNOWLEDGE
The construction industry is organized in projects and current production theory and practice are heavily influenced by the concepts and techniques of project management. According to PMI’s A Guide to the Project Management Body of Knowledge, “a project is a temporary endeavor undertaken to produce a unique product or service.” The making (i.e., manufacturing) of multiple copies of a product does not occur through projects so understood. This focus on product uniqueness and the project form of organization has dominated thinking about production of the built environment so far as to discourage learning from non-project industries such as product manufacturing (Koskela, 1992).
Again according to PMI (1996), project management includes the management of integration, scope, time, cost, quality, human resources, communications, risk, and procurement. Any or all of these could conceivably concern the actual production process itself, but perhaps most of all time and cost.
Time management is said to consist of activity definition, activity sequencing, activity duration estimating, schedule development, and schedule control. The focus is entirely on delivering project objectives; in Koskela’s terms, on the transformation or conversion processes (activities) and not on flow or value generation processes. Activities are to be defined so as to facilitate a division of labor and subsequent tracking (accounting) of conformance to requirements. There is no mention of structuring work for flow or of defining activities so that they facilitate the actual performance of the work. Activity sequencing assumes that handoffs from one set of specialists to the next occur only once; that there is no repetition or cycling to be managed (“conditional diagramming methods” are mentioned-see page 63-but not developed). Schedule control is concerned with managing changes to the schedule rather than with execution of scheduled work; with the exception of expediting as a type of time management corrective action (see page 72). Cost management is treated very much in the same way as time management. The question for project management thus remains: ‘Who manages production and how?’
PMI differentiates between project processes and product-oriented processes (page 27), the former being characteristic of all types of projects and the latter specific to the various types of production with which projects may be involved. What is missing in this distinction is the concept of the project itself as a temporary production system linked to other temporary and permanent production systems for materials, equipment, labor, etc. Projects as such have no necessary connection with production. For example, a project may be to solve a problem of getting voters to register. In this broad sense of the term, ‘project’ becomes virtually synonymous with a single instantiation of the problem solving process, and project management consists of the tools and techniques for managing
production itself takes place alongside project management, but is not directly the business of project management. Consequently, project control consists of monitoring progress toward project objectives and taking corrective action when the ship appears to be off course.
This concept of project control is very different from production control, which is dedicated to causing events to conform to plan and to replanning when events cannot be conformed. Production control conceives production as a flow of materials and information among cooperating specialists, dedicated to the generation of value for customer and stakeholders.
2.2.2 CRITIQUE OF THE TRADITIONAL PROJECT CONTROL MODEL
Project control has been hitherto conceived and carried out consistently with the conversion or transformation view of projects (Koskela and Huovila, 1997). The received wisdom regarding AEC project control systems is founded on a widely shared conception of their purpose. “This (project control) system must provide the information needed for the project team and project participants to identify and correct problem areas and, ultimately, to keep project costs and schedule ‘under control’.” (Diekmann and Thrush, 1986). The objective is to detect negative variances from target, so corrective action can be taken. This is quite different from the active concept of control dominant in manufacturing production control systems, especially those employing a pull strategy, in which the purpose of control is to cause events to conform to plan. In the following, we further examine traditional project controls and their difference from the concept of control in the Last Planner system, which is to be introduced in Chapter 3.
In traditional project control, the objects of control are time and resources. Resources (labor hours, material, equipment, indirects) are planned and controlled
through cost control systems, the objective of which is productivity, i.e., efficient use of resources. A budget is prepared for each resource, the use of resources is monitored against their budgets, and periodic forecasts are made of resource requirements based on the current state of the project.
Controlling time involves planning, scheduling, and monitoring. Planning decides what is to be accomplished and in what sequence. Scheduling determines task duration and timing. Monitoring checks progress of tasks against the schedule and forecasts when work will be completed. The objective of time control is production or progress, not productivity.
Decisions made regarding budget and schedule, productivity and production must recognize their interdependence. Productivity and production are formally related in earned value systems, which propose a solution to the problem that progress and expenditure of resources need not coincide. Rates of resource consumption are established for the various kinds of work to be performed on a project; e.g., 9.32 engineering labour-hours per piping isometric drawing or 12.4 labour-hours per purchase order. Completing an individual piping isometric drawing earns 9.32 labour-hours regardless of the actual number of hours consumed in its production. Progress toward project completion is tracked by accumulating the earned hours and comparing that to the total hours to be earned for the entire project. For example, suppose the project schedule calls for production of 10 piping isometric drawings at time t, but only 9 drawings have been produced. Only 83.88 (9 x 9.32) hours have been earned of the 93.2 scheduled, so that portion of the project is 10% behind schedule (83.88/93.2=.90). That is a measure of production against schedule.
Productivity can be quite a different story. Suppose it has taken only 80 hours to produce the 9 piping isometric drawings. Since 83.88 hours were earned, the performance factor is .95 and the piping group is operating at 95% of its budget for isometric drawings. In this case, the project is behind schedule, but under budget. Production is poor and productivity is good.
Earned value analysis is a means for controlling projects through productivity and progress. By itself, it would have the design manager believe that a project is performing well if it is earning labor hours at the budget unit rate and also earning sufficient hours to maintain a scheduled earnings plan expressed as percentages of earned hours to total hours to be earned. The obvious weakness in this control mechanism is that projects may exhibit budget productivity and be on the earnings plan, but not be doing the right work in the right way at the right time. Although things appear to be on track, the train is destined to eventually run off the rails because work is being produced that does not conform to product quality requirements or to process quality requirements (e.g., out of sequence). Consequently, quality control is invoked as a separate control mechanism, although rarely if ever controlling against the objective of expressing customer needs in engineering specifications, but rather controlling against the objectives of avoiding calculational and dimensional errors. As for the issue of the timing of work, it has proven necessary to establish schedule milestones to enforce adherence to a work sequence. These rear guard actions are frequently ineffective against the dominant progress and productivity controls, which consequently cause managers to throw the lever in the wrong direction because they misevaluate actual project performance (Howell and Ballard, 1996).
Work Breakdown Structure (WBS) is a key element in traditional project controls. “A WBS provides a framework for integrated schedule and cost planning and allows for monitoring and control by management by establishing the manner in which estimates are assigned and costs are accumulated and summarized.” (p. 21, Diekmann and Thrush, 1986). The objective is to divide the work to be done in the project into parts so they can be monitored and controlled. No mention is made of the production process as such. [NB: Inclusion of the flow view adds new criteria to the decomposition process. Roughly speaking, we want to break the whole into parts so we can more easily put the parts back together again. Structure work for flow and assembly, not only for budgeting and monitoring.]
Further decomposition in the traditional process eventually defines work packages as the smallest unit. Work packages often correspond to contract packages or to pay items within a single contract. The dominance of the conversion view is perhaps best revealed in the following quotes: “A work package is a cost center.” (p. 73, Neil, James M. Construction Cost Estimating for Project Control, 1982). “The WBS provides the framework for defining the project from the top all the way down to its smallest components and for accumulating the costs associated with each piece. In so doing, the WBS provides a data base from which problem areas can be identified, forecasts made, and corrective action can be taken.” (p. 21, Diekmann and Thrush, 1986). It appears to be assumed that costs arise within that part of the project in which they are detected. Further, control is essentially control of behaviour, given the default assumption that tasks/work packages/contracts can be carried out. The flow view, with its interdependence of parts (both as regards the 'product' and the process of making that
We are clearly dealing here with a type of push system and the controls appropriate to a push system.
Despite the focus on cost and schedule ‘accounting’, theorists recognize the primacy of the control act itself. “Without corrective actions a project control system becomes merely a cost/schedule reporting system.” (p. 29, Diekmann and Thrush, 1986). However, the traditional view is that control consists of correcting deviations from plan. Deviations are expected, but that expectation is not rooted in the idea that variation is natural, but rather that sin is inevitable. Diekmann and Thrush devote less than two pages of a 108 page paper to corrective action and provide no more advice than to inform managers and supervisors at every level in the project about deviations so they can “…correct those trouble spots.” (p. 28). They appear to assume that causes of deviation will be apparent and the appropriate corrective action obvious. “These problems can be easily traced to their source allowing early detection of unfavorable trends.” (p. 33, Diekmann and Thrush, 1986). If the standard corrective actions are indeed ‘Try harder!’ and ‘Add more men!’, that would be consistent with the traditional view.
Advocates of system dynamics have proposed to supplement traditional network analyses and models, adding to the “…growing evidence that network analysis on its own is not sufficient to model and manage the behaviour of projects.” (Williams et al., 1995, p. 154). They propose to provide additional information to project managers so they avoid misevaluating the state of the project and consequently making decisions that cause things to get worse rather than better (See p. 125 of Rodrigues, 1994). Ballard and Howell (1996) suggest that it is impossible to make good decisions about causes or corrections of deviations, relying only on productivity and progress data, without understanding work flow. One can hardly avoid concluding that the traditional control
system is indeed based almost exclusively on the conversion or activity view of the production system.
2.3 Previous Applications of Production Control Concepts to the AEC
Industry
A survey of the literature reveals several primary contributors to the theory and practice of production (as opposed to project) control in the construction industry. Ballard and Howell’s contributions are described in Chapter Three. Melles and Wamelink (1993) developed a very similar line of thinking independently, culminating in their joint PhD thesis at Delft University, The Netherlands. Lauri Koskela, Senior Researcher at Finland’s building research institute, VTT, is the leading theorist in production management in construction. The University of Reading has been active in the field of production management for a number of years. John Bennett’s Construction Management from 1985 is an excellent example of their work. Addis’ 1990 book, Structural engineering: the nature and theory of design, is also a highly relevant work for this research. Alexander Laufer’s work on project planning takes a production control orientation by virtue of its focus on uncertainty and variability and their management. Given the relative obscurity of Melles and Wamelink’s, only their work is presented in detail. The work of Koskela is described only to the extent needed to remind the reader of his vital contributions. That should in no way be taken as an indication of relative importance of the various contributions.
2.3.1 MELLES AND WAMELINK
Introducing their discussion of the theory of production control, Melles and Wamelink (1993) explain, “Contrary to what is customary in the construction industry we shall not
assume, beforehand, the theories in the field of project management. …Production control in construction companies has traditionally been aimed at the control of projects.” For Melles and Wamelink, production control consists of “…the activities relating to the adjustment of all aspects of the production process, so that the preconditions in which the production process is to be executed, are met.” Drawing on manufacturing production control, they emphasize: 1) Thinking in terms of hierarchical levels of decision; i.e., control at company level, factory level, and production unit level, and 2) Thinking in terms of decision functions within the hierarchical levels; i.e., aggregate production control, material coordination, workload control, workorder release, workload acceptance, detailed workorder scheduling, capacity allocation, and shop floor control. The manufacturing model on which they rely is that of Bertrand et al., 1990.
Melles and Wamelink propose a ‘translation’ of the manufacturing model into decision functions appropriate to various types of construction, identifying at the ‘factory’ level project coordination (achieved in part by network schedules), mobilisation planning (by means of “six weeks scheme”), and allocation planning (by means of “task scheme”).
In addition to the primary contribution of directing attention to manufacturing theory and practice, Melles and Wamelink’s work identifies functionalities AEC industry production control systems should possess. Their specific objective was to assist in the design of information systems. Consequently, they did not explicitly apply their model to evaluation of current management systems and practice. However, the overwhelmingly negative results of so doing are implicit in their critique of project management software. For example, speaking of project coordination, they comment, “…it can immediately be
deduced that the project management software available on the market is indeed about a certain aspect (within the framework, the decision function project coordination). The other decision functions (resource planning, mobilization planning, etc.) are, generally speaking, not recognizable.” (p. 35). This critique is made more explicitly in Wamelink et al., 1993.
2.3.2 KOSKELA
Lauri Koskela (1999) proposes the following design criteria or principles for a production control system. In fact, he claims they are true for the Last Planner system, which is to be presented in Chapter Three:
"The first principle is that the assignments should be sound regarding their prerequisites. This principle has also been called the Complete Kit by Ronen (Ronen 1992). The Complete Kit suggests that work should not start until all the items required for completion of a job are available. Thus, this principle strives to minimize work in suboptimal conditions.
"The second principle is that the realization of assignments is measured and monitored. The related metrics, Percent Plan Complete (PPC), is the number of planned activities completed, divided by the total number of planned activities, and expressed as a percentage. This focus on plan realization diminishes the risk of variability propagation to downstream flows and tasks.
"Thirdly, causes for non-realization are investigated and those causes are removed. Thus, in fact, continuous, in-process improvement is realized.
"The fourth principle suggests maintaining a buffer of tasks which are sound for each crew. Thus, if the assigned task turns out to be impossible to carry out, the crew
production (due to starving) or reduced productivity (due to suboptimal conditions).
"The fifth priciple suggests that in lookahead planning (with time horizon of 3-4 weeks), the prerequisites of upcoming assignments are actively made ready. This, in fact, is a pull system that is instrumental in ensuring that all the prerequisites are available for the assignments. On the other hand, it ensures that too great material buffers do not emerge on site.”
2.4 Criteria for a Design Production Control System
The preceding review and critique of the literature suggests the following guidelines and criteria for an effective design production control system:
q Variability must be mitigated and remaining variability managed. Variability is virtually
disregarded in current control systems. But the construction industry certainly has its share of variability: variability in quality, variability in processing times, variability in deliveries, etc. Neglect of variability causes greater variability, and there is always an associated penalty. According to Hopp and Spearman (1996), variability results in some or all of the following:
§ buffering of flows, which increases lead times and work-in-process § lower resource utilization
§ lost throughput
q Assignments are sound regarding their prerequisites.
q The realization of assignments is measured and monitored.
q Causes for failing to complete planned work are investigated and those causes are removed.
q A buffer of sound assignments is maintained for each crew or production unit.
q The prerequisites of upcoming assignments are actively made ready.
q The traditional schedule-push system is supplemented with pull techniques. Not only do pull
especially needed in conditions of variability.
q Production control facilitates work flow and value generation. Production thinking and practice in
all areas has focused primarily on the task goals of production and neglected flow and value (Huovila and Koskela, 1997). The object of traditional project control has been behavior. What needs to be controlled is work flow.
q The project is conceived as a temporary production system. The model for corrective action in
traditional project control is course correction, drawn by analogy from the path of a vehicle bound for a specific destination with a target arrival time and a specified spending budget or otherwise limited resources. If the project is to be conceived rather as a temporary production system, the course correction model is radically oversimplified and inappropriate. The flow of materials and information is what is to be controlled. They flow through very complex networks of temporary and permanent production systems. Corrective action must be taken within an understanding of these networks and of the impact of changes in sequence, processing methodologies, buffer location and sizing, local control strategies (e.g., pull or push), etc.
q Decision making is distributed in production control systems. Traditional project control assumes
the necessity and possibility of central control. The underlying image is that of a single mind and many hands. Arguably, dynamic production systems cannot be controlled centrally, but rather are adaptive creatures driven by decision making at their periphery.
q Production control resists the tendency [of designers and engineers] toward local suboptimization
(Green, 1992). Green's comment was specifically directed to the tendency of designers and
engineers toward local suboptimization, but that is a general tendency of any system in which there is a division of labor.
In Chapter Three, the Last Planner system of production control is described and evaluated against these criteria.
CHAPTER THREE: DESCRIPTION AND HISTORY OF
THE LAST PLANNER SYSTEM OF PRODUCTION
CONTROL
3.1 Hierarchical Structure
Aside from the simplest and smallest jobs, design and construction require planning and control done by different people, at different places within the organization, and at different times during the life of a project. Planning high in the organization tends to focus on global objectives and constraints, governing the entire project. These objectives drive lower level planning processes that specify means for achieving those ends. Ultimately, someone (individual or group) decides what physical, specific work will be done tomorrow. That type of plans has been called "assignments". They are unique because they drive direct work rather than the production of other plans. The person or group that produces assignments is called the "Last Planner" (Ballard and Howell 1994).
3.2 Should-Can-Will-Did
The term "assignments" stresses the communication of requirements from Last Planner to design squad or construction crew. But these products of planning at the production unit level are also commitments to the rest of the organization. They say what WILL be done, and (hopefully) are the result of a planning process that best matches WILL with SHOULD within the constraints of CAN.
Figure 3.1
SHOULD
CAN LAST PLANNER PLANNING WILL
PROCESS
The formation of assignments in the Last Planner planning process.
Unfortunately, last planner performance is sometimes evaluated as if there could be no possible difference between SHOULD and CAN. "What will we do next week?” “Whatever is on the schedule," or “Whatever is generating the most heat.” Supervisors consider it their job to keep pressure on subordinates to produce despite obstacles. Granted that it is necessary to overcome obstacles, that does not excuse creating them or leaving them in place. Erratic delivery of resources such as input information and unpredictable completion of prerequisite work invalidates the presumed equation of WILL with SHOULD, and quickly results in the abandonment of planning that directs actual production.
Failure to proactively control at the production unit level increases uncertainty and deprives workers of planning as a tool for shaping the future. What is needed is to shift the focus of control from the workers to the flow of work that links them together. The Last Planner production control system is a philosophy, rules and procedures, and a set of tools that facilitiate the implementation of those procedures. Regarding the procedures, the system has two components: production unit control and work flow
control. The job of the first is to make progressively better assignments to direct workers through continuous learning and corrective action. The function of work flow control is perhaps evident in its name—to proactively cause work to flow across production units in the best achieveable sequence and rate.
3.3 Production Unit Control
The key performance dimension of a planning system at the production unit level is its output quality; i.e. the quality of plans produced by the Last Planner. The following are some of the critical quality characteristics of an assignment:
q The assignment is well defined.
q The right sequence of work is selected.
q The right amount of work is selected.
q The work selected is practical or sound; i.e., can be done.
“Well defined” means described sufficiently that it can be made ready and completion can be unambiguously determined. The "right sequence" is that sequence consistent with the internal logic of the work itself, project commitments and goals, and execution strategies. The "right amount" is that amount the planners judge their production units capable of completing after review of budget unit rates and after examining the specific work to be done. "Practical" means that all prerequisite work is in place and all resources are available.
The quality of a front line supervisor's assignments may be reviewed by a superior prior to issue, but such in-process inspection does not routinely produce measurement data, even when corrections are necessary. Planning system performance is more easily measured indirectly, through the results of plan execution.
by the total number of planned activities, expressed as a percentage. PPC becomes the standard against which control is exercised at the production unit level, being derivative from an extremely complex set of directives: project schedules, execution strategies, budget unit rates, etc. Given quality plans, higher PPC corresponds to doing more of the right work with given resources, i.e. to higher productivity and progress.
Percent Plan Complete measures the extent to which the front line supervisor's commitment (WILL) was realized. Analysis of nonconformances can then lead back to root causes, so improvement can be made in future performance. Measuring performance at the Last Planner level does not mean you only make changes at that level. Root causes of poor plan quality or failure to execute planned work may be found at any organizational level, process or function. PPC analysis can become a powerful focal point for breakthrough initiatives.
The first thing needed is identification of reasons why planned work was not done, preferably by front line supervisors or the engineers or craftsmen directly responsible for plan execution. Reasons could include:
q Faulty directives or information provided to the Last Planner; e.g. the information system incorrectly indicated that information was available or that prerequisite work was complete.
q Failure to apply quality criteria to assignments; e.g. too much work was planned.
q Failure in coordination of shared resources; e.g. lack of a computer or plotter.
q Change in priority; e.g. workers reassigned temporarily to a "hot" task.
q Design error or vendor error discovered in the attempt to carry out a planned activity.
This provides the initial data needed for analysis and improvement of PPC, and consequently for improving project performance.
3.4 Work Flow Control
production units in a desired sequence and rate. Production Unit Control coordinates the execution of work within production units such as construction crews and design squads. Work Flow Control coordinates the flow of design, supply, and installation through production units.
In the hierarchy of plans and schedules, the lookahead process has the job of work flow control. Lookahead schedules are common in current industry practice, but typically perform only the function of highlighting what SHOULD be done in the near term. In contrast, the lookahead process within the Last Planner system serves multiple functions, as listed in Table 3.1. These functions are accomplished through various specific
processes, including activity definition, constraints analysis, pulling work from upstream production units, and matching load and capacity, each of which will be discussed in the following pages.
Figure 3.2
5 Selecting,
sequencing, & sizing work we think can be done
Master Schedule Make work ready by screening & pulling Information Selecting, sequencing, & sizing work
we know can be done Current status
& forecasts Lookahead
Workable Backlog Weekly Work Plans Production Resources Completed Work Chart PPC & Reasons Action to prevent repetitive errors
PLANNING SYSTEM
Last Planner System with Lookahead Process highlighted
The vehicle for the lookahead process is a schedule of potential assignments for the next 3 to 12 weeks. The number of weeks over which a lookahead process extends is decided based on project characteristics, the reliability of the planning system, and the lead times for acquiring information, materials, labor, and equipment. Tables 3.2 and 3.3 are examples of construction and engineering lookahead schedules, respectively. The lookahead schedule is not a simple drop out from the master schedule. Indeed, it is often beneficial to have the team that is to do the work in the next phase of a project collectively produce a phase schedule that serves to coordinate actions that extend beyond the lookahead window (the period of time we choose to look ahead).
Table 3.1
Functions of the Lookahead Process
• Shape work flow sequence and rate
• Match work flow and capacity
• Decompose master schedule activities
into work packages and operations
• Develop detailed methods for executing
work
• Maintain a backlog of ready work
• Update and revise higher level schedules
as needed.
Functions of the Lookahead Process
Prior to entry into the lookahead window, master schedule or phase schedule activities are exploded into a level of detail appropriate for assignment on weekly work plans, which typically yields multiple assignments for each activity. Then each assignment is subjected to constraints analysis to determine what must be done in order to make it ready to be executed. The general rule is to allow into the lookahead window, or allow to advance from one week to the next within the lookahead window, only activities that can be made ready for completion on schedule. If the planner is not confident that the constraints can be removed, the potential assignments are retarded to a later date.
Figure 3.3 is a schematic of the lookahead process, showing work flowing through time from right to left. Potential assignments enter the lookahead window 6 weeks ahead of scheduled execution, then move forward a week each week until they are allowed to enter into workable backlog, indicating that all constraints have been removed and that they are in the proper sequence for execution. If the planner were to discover a
constraint (perhaps a design change or acquisition of a soils report) that could not be removed in time, the assignment would not be allowed to move forward. The objective is to maintain a backlog of sound work, ready to be performed, with assurance that everything in workable backlog is indeed workable.13 Weekly work plans are then formed from workable backlog, thus improving the productivity of those who receive the assignments and increasing the reliability of work flow to the next production unit.
Table 3.2
PROJECT: Pilo t 5 WK LOOKAHEAD
ACTIVITY 3 / 9 # # # NEEDS
M T W T F S M T W T F S M T W T F S M T W T F S
Sc ot t 's cre w
"CUP" AHUs-1 0 CHW, 2 HW X X X X X X X X X X X X X X X CHW d elive rs 1 -8 -9 7 t hru
1- 13 .HW de live rs 1 -2 0.
Punch, lab e l, & t ag AHUs x x x Mat er ials on sit e
Ro n's cre w
DI St e a m t o Humid ifier x x x Mat er ials on sit e
DI St e a m Blowd own x x Ch e ck ma t e rial
DI St e a m Co nd. t o x x x x x x x x x x x x x Mat er ial on s it e
co ole rs ( 1 3 ) Charles ' cre w
2 00 de g HW 1 -"H" x x x Mat l d e live ry 1 -8 -9 7 2 0 0 de g HW 1 -" B" x x x x x x x x x x Re lea se mat l for 1 -1 5 -9 7 & 1 -" D"
1 st flr 2 0 0 d eg HW x x x x x x x x x x Mat er ial on s it e . Nee d We s t
guide s & ancho rs Wing flr co ve re d .
Richa rd's c rew
2 -"A" HW & CHW x x x x x Co nt rol va lve s for adde d VAV co ils
CHW in C- E-G t unn els x x x x x x x x x x x x x x x Ne e d t u nne ls paint e d & re le as e ma t e rials
Misc FCUs & co nd. drains x x x x x x x x x x Ta ke of f & o rde r ma t e rials
in "I" , " J", & "K" 1 st flr
Punch, lab e l & t ag x x x x x x x x x x Mat er ial on s it e 1 / 1 3/ 97 1 / 20 / 9 7 1 / 2 7/ 97 2 / 3 / 9 7
Construction Lookahead Schedule14
13
Deliberately building inventories, inventories of ready work in this case, may seem contradictory to the goals of just-in-time. To clarify, inventories of all sort are to be minimized, but as long as there is variability in the flow of materials and information, buffers will be needed to absorb that variability. Reducing variability allows reduction of buffer inventories.
3.4.1 CONSTRAINTS ANALYSIS
Once assignments are identified, they are subjected to constraints analysis. Different types of assignments have different constraints. The construction example in Table 3.4 lists contract, design, submittals, materials, prerequisite work, space, equipment, and labor; plus an open-ended category for all other constraints. Other constraints might include permits, inspections, approvals, and so on. Design constraints can virtually be read from the Activity Definition Model: clarity of directives (level of accuracy required, intended use of the output, applicable section of code), prerequisite work (data, evaluations, models), labor and technical resources. We previously met these constraints in the discussion of Production Unit Control; then as reasons for failing to complete assignments on weekly work plans.
Table 3.3 Project: Discipline: Process Planner: s Checked By; x Prep. Dt: 3/14/02
Week Ending: Week Ending: Week Ending: Week Ending:
Activity 3/28/02 4/4/02 4/11/02 4/18/02 OUTSTANDING NEEDS
M T W T F M T W T F M T W T F M T W T F
Provide const support (Q
& A) x x x x x x x x x x x x x x x x x x
Need questions from subs. Review submittal(s) x x Need submittals from sub. Aid with tool install dsgn
effort. x x x x x x x x x x x x x x x x x x x
Frozen layout, pkg 1 dwgs. Design drains from tools
to tunnel tie-ins. x x x
Frozen layout, input from tool install on installation preferences
Help layout people complete a layout that will work well with tool install routing and drains into the tunnel.
x x Correct tool list.
Complete Pkg 2 specifications
x x x x x Final eqpt and mtl usage from mech & tool install.
Create work plans x x x x
Send package to QA/QC reviewer for drain design review
x x Final design dwgs for drains; plot time Start/complete QA/QC
review
x x Set of Package 2 review docs, dwgs
Engineering Lookahead Schedule
Constraints analysis requires suppliers of goods and services to actively manage their production and delivery, and provides the coordinator with early warning of problems, hopefully with sufficient lead time to plan around them. In the absence of constraints analysis, the tendency is to assume a throw-it-over-the-wall mentality; to become reactive to what happens to show up in your in-box or laydown yard.
Figure 3.3
6
Screen assignments & make ready each week enough work to maintain 2 week
workable backlog Notify coordinator of constraints status Explode scheduled activities into work packages on entry to the lookahead window 1 2 3 4 5 Assign-ments Workable Backlog
Master schedule activities entering 6th week
Reasons why planned work not completed
The Lookahead Process: Make Ready by
Screening & Pulling
Make Ready by Screening and Pulling
3.4.2 PULLING
Pulling is a method of introducing materials or information into a production process. The alternative method is to push inputs into a process based on target delivery or completion dates. Construction schedules have traditionally been push mechanisms,
seeking to cause intersections in the future of interdependent actions. Table 3.4 I D A c t i v i t y S t a r t C o n t r a c t D e s i g n S u b m i t t a l s M a t e r i a l s P r e - R e q u i s i t e S p a c e E q u i p m e n t L a b o r O t h e r 2 6 0 S m a l l I n t e r i o r W a l l F o r m s L i n e s 4 M . 8 , 3 M , 3 -K , 4 - -K . 8 , 3 - H 2 / 9 / 9 8 O K R F I 6 8 O K O K r e b a r O K O K O K N o n e 3 1 0 L a r g e In t e r i o r W a l l L i n e L F o r m 2 / 9 / 9 8 7 0 0 In t e r i o r S m a l l W a l l s 3 F a n d 3 D F o r m s 2 / 9 / 9 8 1 1 4 2 S m a l l I n t e r i o r W a l l F o r m s L i n e s 5 - M . 8 , a n d 5 - K . 8 2 / 9 / 9 8 1 7 0 E a s t W a l l B e t w e e n L i n e s 2 a n d 6 L i n e D o u b l e U p 2 / 1 3 / 9 8 7 2 0 In t e r i o r S m a l l W a l l s 3 F a n d 3 D D o u b l e - u p 2 / 1 3 / 9 8 1 1 4 6 S m a l l I n t e r i o r W a l l s L i n e s 5 M . 8 , a n d 5 -K . 8 D o u b l e - u p 2 / 1 3 / 9 8 3 2 2 L a r g e In t e r i o r W a l l L i n e L D o u b l e u p 2 / 1 6 / 9 8 2 9 0 S m a l l I n t e r i o r W a l l s L i n e s 4 M . 8 , 3 M , 3 K , 4 -K . 8 , 3 - H D o u b l e - u p 2 / 1 7 / 9 8 7 3 5 In t e r i o r S m a l l W a l l s 3 F a n d 3 D S t r i p 2 / 1 8 / 9 8
Screening Assignments: Statusing
Constraints
Constraints Analysis
By contrast, pulling allows materials or information into a production process only if the process is capable of doing that work. In our Last Planner system, conformance of assignments to quality criteria constitute such a check on capability. Further, making assignments ready in the lookahead process is explicitly an application of pull techniques. Consequently, Last Planner is a type of pull system.
Figure 3.4
A Traditional (Push)
Planning System
PLANNING THE WORK INFORMATION PROJECT OBJECTIVES SHOULD EXECUTING THE PLAN RESOURCES DIDA Traditional (Push) Planning System
Certain things have long been pulled as opposed to pushed; e.g., concrete. With its short shelf life, concrete cannot be ordered too far in advance of need. Fortunately, the lead time15 for concrete is short, so it is usually possible to wait until you know when it will be needed before ordering it.
Generally, a window of reliability greater than supplier lead time is needed in order for pulling to be most effective. Otherwise, the pulled items may not match up with the work to which they are to be applied. In the industry now, supplier lead times are for the most part much greater than our accurate foresight regarding work completion, hence perhaps a reason for the infrequent use of pulling mechanisms.
Figure 3.5
Last Planner: A Pull System
SHOULD
CAN WILL
LAST PLANNER PLANNING
PROCESS
Last Planner: A Pull System
3.4.3 MATCHING LOAD AND CAPACITY
Matching load to capacity within a production system is critical for productivity of the production units through which work flows in the system, and is also critical for system cycle time, the time required for something to go from one end to the other.
Along with its other functions, the lookahead process is supposed to maintain a backlog of workable assignments for each production unit (PU). To do so requires estimating the load various chunks of work will place on PUs and the capacities of PUs to process those chunks of work. Current estimating unit rates, such as the labor hours required to erect a ton of steel, are at best averages based on historical data, which are themselves laden with the tremendous amounts of waste imbedded in conventional practice. When
